One API for every
quantum computer.
Submit OpenQASM 3 over REST. Pauli quotes your circuit on every device, then routes by price, queue, or fidelity.
curl https://pauli.xyz/v1/jobs \
-H "Authorization: Bearer $PAULI_API_KEY" \
-d '{
"program": {"format": "openqasm3",
"source": "OPENQASM 3.0; include \"stdgates.inc\"; qubit[2] q; bit[2] c; h q[0]; cx q[0], q[1]; c = measure q;"},
"shots": 1000,
"routing": {"policy": "cheapest", "allow_simulator_fallback": true},
"billing": {"max_spend_usd": "5.00"}
}'HTTP/2 202
{
"job_id": "7b1c2e90-4f3a-4c2e-9b1a-2f6f6f0e8d21",
"status": "queued",
"routing_decision": {
"selected_device": "ionq.qpu.forte-enterprise-1",
"score": 0.8121,
"reason": "cheapest policy; est. fidelity 0.953; cost $1.90; predicted queue 6871s; connectivity fit 1.00 (no SWAP overhead)"
},
"cost_estimate": {"amount": "1.900000", "unit": "ionq_aqt"}
}The quantum cloud is a market with no price tags.
An independent study ran 5,000+ jobs (~$600k) across the major quantum clouds:
cost spread for the same workload: a 10-qubit, 500-shot QFT ran $1.03 on IQM Garnet, $2,289.86 on Quantinuum H-Series.
of vendor queue estimates overshot the real wait.
the same physical QPU (IonQ Aria-1), priced per storefront.
Source: “Three Months in the Life of Cloud Quantum Computing,” arXiv:2601.09943 (Jan 2026).
Compare the fleet before you spend a cent.
Live prices, queues, and fidelity, from the same transpile the router scores with.
| Device | Status | Est. price | Predicted queue | Est. fidelity |
|---|---|---|---|---|
| Loading live estimates… | ||||
Your own circuits get the same comparison, authenticated, via
POST /v1/jobs/estimate.